Implications of variability in cell membrane permeability for design of methods to remove glycerol from frozen-thawed erythrocytes

In North America, red blood cells (RBCs) are currently cryopreserved in a solution of 40% glycerol. While glycerol is not inherently toxic to humans, it must be removed prior to transfusion to prevent intravascular osmotic hemolysis. The current deglycerolization procedure requires about 45 min per RBC unit. We previously presented predictions suggesting that glycerol could be safely removed from RBCs in less than 1 min. However, experimental evaluation of these methods resulted in much higher hemolysis than expected. Here we extend our previous study by considering both concentration-dependence of permeability and variability in permeability values in the mathematical optimization algorithm. To establish a model for the concentration dependence of glycerol permeability, we combined literature data with new measurements of permeability in the presence of 40% glycerol. To account for cell-dependent variability we scaled the concentration-dependent permeability model to define a permeability range for optimization. Methods designed using a range extending to 50% of the model-predicted glycerol permeability had a duration of less than 3 min and resulted in hemolysis ranging from 34% to 83%; hemolysis values were highly dependent on the blood donor. Extending the permeability range to 5% of the model-predicted value yielded a 30 min method that resulted in an average hemolysis of 12%. Our results suggest high variability in the glycerol permeability between donors and within a population of cells from the same donor. Such variability has broad implications for design of methods for equilibration of cells with cryoprotectants.